Sign in

my elasticsearch node gets oomkilled at random. memory graphs look fine! !!

dockerkubernetesmemory46 replies · 548 views · started
Reply
  1. Silas Weber
    @silas.weber

    respectfully disagree, i tried that and it made 256mi worse idk maybe im wrong lol

    0 2 helpful
  2. Mila
    @meriksen

    ahh classic runtimes need to know about the cgroup limit. Modern JVMs respect it, but set heap explicitly with -XX:MaxRAMPercentage=75 so heap plus metaspace, threads and buffers fit. for Node.js, set --max-old-space-size below teh limit. the rest of the container needs headroom beyond the heap.

    1 0 helpful
  3. Bram Ueda
    @bueda

    btw lmao this is better than netflix

    1 0 helpful
  4. Ash
    @kernel_shed

    @kernel_shed did u ever fix it? ?

    3 0 helpful
  5. Diego Pereira
    @diegop96

    ahh classic honestly just reset teh default config and save yourself the pain.. 🙏 exact same thing happened to my homelab box last month worked for me anyway 🤷

    12 0 helpful
  6. Silas Weber
    @silas.weber
    in reply to @kernel_shed

    @silas.weber cool story. still wrong tho idk maybe im wrong lol

    @kernel_shed every single time someone suggests fixing it properly the thread gets longer and nothing gets fixed just my 2 cents

    2 0 helpful
  7. Oskar Quintero
    @oskar_heap

    yeahhh dashboards often sample every 30-60 sceonds and show averages. A short spike (a big request, a batch load) can blow past 768Mi between samples. Look at max, not average, and at container_memory_working_set_bytes rather than RSS alone. hope that helps :)

    0 1 helpful
  9. Silas Weber
    @silas.weber

    oof have u checked your memory settings? mine were reset after an update

    thumbsup

    5 0 helpful
  11. Amara Holm
    @aholm

    not gonna lie this is driving me nuts spent a whole week on elasticsearch and im still not done im about to throw this thing out the window lol

    2 0 helpful
  12. Joon H.
    @jhalvorsen

    so annoyed rn my office machine is officially haunted thx in advance!!

    7 0 helpful
  13. Ash
    @kernel_shed

    this. sooo much this

    6 5 helpful
  14. Rhea
    @daemonowl

    oof +1 same here hope that helps :)

    2 0 helpful
  15. Ash
    @kernel_shed
    in reply to @silas.weber

    @kernel_shed every single time someone suggests fixing it properly the thread gets longer and nothing gets fixed just my 2 cents

    @silas.weber fine. agree to disagree. (ur still wrong)

    10 1 helpful
  16. Silas Weber
    @silas.weber
    in reply to @kernel_shed

    @silas.weber fine. agree to disagree. (ur still wrong)

    @kernel_shed lol ok, keep believing fixing it properly is fine idk maybe im wrong lol

    20 2 helpful
  17. Ash
    @kernel_shed
    in reply to @silas.weber

    @kernel_shed lol ok, keep believing fixing it properly is fine idk maybe im wrong lol

    @silas.weber ive been doing this for 15 years, u can stop explaining a clean reisntall to me !!! idk maybe im wrong lol

    5 2 helpful
  18. Silas Weber
    @silas.weber
    in reply to @kernel_shed

    @silas.weber ive been doing this for 15 years, u can stop explaining a clean reisntall to me !!! idk maybe im wrong lol

    @kernel_shed did u even read my post or just teh first line 😤 idk maybe im wrong lol

    6 4 helpful
  19. Ash
    @kernel_shed
    in reply to @silas.weber

    @kernel_shed did u even read my post or just teh first line 😤 idk maybe im wrong lol

    @silas.weber not this again 🙄 lol just my 2 cents

    0 0 helpful

Reply to this topic

Sign in or create an account to reply.